Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Weather image intelligent identification method and system based on multi-depth convolutional neural network fusion

A convolutional neural network and weather image technology, applied in the field of weather image intelligent recognition methods and systems, can solve the problems of ineffective guarantee of deep learning network training data quality and inefficient image features, etc.

Active Publication Date: 2020-07-31
CENT SOUTH UNIV
View PDF7 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

It solves the problems of inefficiency in manually extracting image features in existing traditional weather recognition methods and the inability to effectively guarantee the quality of deep learning network training data, and effectively improves the recognition accuracy of weather images

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Weather image intelligent identification method and system based on multi-depth convolutional neural network fusion
  • Weather image intelligent identification method and system based on multi-depth convolutional neural network fusion
  • Weather image intelligent identification method and system based on multi-depth convolutional neural network fusion

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0068] The overall implementation process of a weather image intelligent recognition method based on multi-depth convolutional neural network fusion is as follows: figure 1 As shown, proceed as follows:

[0069]Step A: Collect images of 9 weather phenomena including rime, rime, rain, snow, hail, dew, frost, fog or haze, and icing, and perform data cleaning, data amplification, and standardization processing on weather images, and formulate them as For a data set in a certain format, the overall steps are as follows figure 2 shown.

[0070] 1) Data Analysis

[0071] According to the collected 9,000 images of nine different weather conditions, the data set is constructed, and the image labels are 1-9, corresponding to rime, rime, rain, snow, hail, dew, frost, fog or haze, and icing. All the data is divided into training set and test set according to the ratio of 2:1.

[0072] Analyzing the data, the distribution of the number of images with different labels in the training ...

Embodiment 2

[0111] Automatic recognition of individual weather images. The first step is the construction of the training data set, collecting 6,000 images of 9 weather phenomena related to rime, rime, rain, snow, hail, dew, frost, fog or haze, and icing. First, analyze the data and perform data cleaning. Class balancing, data augmentation, and normalization to a dataset of size 512×512.

[0112]In the second step, four deep convolutional neural network models are selected: ResNe152, DenseNet169, ResNext101_64x4d, SE_ResNeXt101_32x4d, modify the fully connected layer of the network, and add a 27×1 fully connected layer before the output layer to form the feature layer of the network. The loss function of the network is shown in formula (1), and its weight parameters are set according to Table 1. The built network needs to be trained first. It is recommended to use GPU to accelerate the training process. The training batch size is 32. The optimizer is Adam and the initial learning rate is...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a weather image intelligent identification method and system based on multi-depth convolutional neural network fusion, and the method comprises the following steps: A, collecting an image, and carrying out the preprocessing of the image; B, establishing four deep convolutional neural network models with different structures, respectively improving full connection layers ofthe deep convolutional neural network models, adding a feature layer, and performing network training based on the high-quality training data obtained in the step A; C, extracting features of newly added feature layers of the four obtained deep learning models, and performing training by adopting an Xgboost ensemble learning model to obtain a fusion model; and D, amplifying a weather image to be identified, identifying the amplified weather image by the obtained fusion model, and performing voting to obtain an identification type with the highest final vote number. According to the method, theproblem of low efficiency of manually extracting the image features in an existing traditional weather recognition method is solved, and the recognition accuracy of the deep learning model is effectively improved.

Description

technical field [0001] The invention belongs to the field of weather image recognition, in particular to a method and system for intelligent recognition of weather images based on multi-depth convolutional neural network fusion. Background technique [0002] Weather conditions are always accompanied by people's daily life, and have a profound impact on people's basic necessities of life. For example, people pay special attention to traffic safety in rainy, snowy and foggy weather, and they need to wear masks or avoid traveling when the smog is severe. People make daily travel arrangements and daily activities based on weather conditions. Traditional weather recognition generally requires a large number of sensors for data collection and relies on a large number of manual observations by professional meteorologists, and the cost and efficiency are greatly limited. With the continuous development of computer technology and deep learning, image recognition technology has been...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06T3/40G06T3/60G06T5/00G06T5/50G06T7/11G06N3/08G06N3/04G06K9/62
CPCG06N3/084G06T7/11G06T5/50G06T3/40G06T3/60G06T2207/20132G06N3/045G06F18/213G06F18/24323G06F18/214G06T5/73G06T5/70
Inventor 郭璠李伟清唐琎邱俊峰
Owner CENT SOUTH UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com